The new study, jointly prepared by the ICCT, the Netherlands’ Organisation for Applied Scientific Research (TNO), and Germany’s Institut für Energie- und Umweltforschung Heidelberg (IFEU), describes the increasing real-world efficiency gap using systematic statistical analysis.

A technical definition of real-world driving is elusive because of variations in vehicle types, driving behavior, and driving conditions. Nonetheless, in aggregating data on almost 600,000 vehicles from eleven data sources and six countries, this study reveals a clear trend over time: the divergence (or “gap”) between real-world and official CO₂ emissions increased from approximately 8 percent in 2001 to 40 percent in 2014. Each data source used for the study includes a unique set of vehicles and drivers, so estimates of the divergence of real-world from official values vary among them. however, the increase in the gap cannot be explained based on driving behavior or differences between the data sources, but is instead a result of increasingly unrealistic type-approval values.

—“From laboratory to road: A 2015 update”

Divergence between real-world and manufacturers’ type-approval CO2 emissions for various real-world data sources, including average estimates for private cars, company cars, and all data sources. Source: ICCT. Click to enlarge.

The analysis draws on data from a number of different sources: the user websites spritmonitor.de (Germany) and honestjohn.co.uk (United Kingdom); the leasing companies Travelcard (Netherlands) and LeasePlan (Germany); the car and consumer magazines Auto Bild (Germany), auto motor sport (Germany and Sweden), WhatCar? (United Kingdom), km77.com (Spain) and the car club TCS (Switzerland).

The researchers assessed in detail the underlying reasons for the growing divergence, and identified four key factors:

• Chassis dynamometer testing. Under the EU regulation, there are a number of “loopholes” that can potentially be exploited by vehicle manufacturers during chassis dynamometer testing. These include break-in periods for the test vehicle, tolerances regarding laboratory instruments, the state of charge of the vehicle’s battery, special test driving techniques, and use of pre-series parts that are not representative of production vehicles. The ICCT analysis indicates that vehicle manufacturers have found ways to optimize chassis dynamometer type-approval testing over time, which at the same time made it less representative of average real-world driving conditions. As a result, the impact of chassis dynamometer testing flexibilities on the divergence between type-approval and real-world CO₂ emissions today is estimated to explain more than half of the overall divergence observed.

• Road load determination. Coefficients for road load are used to characterize the forces (mainly aerodynamic drag and rolling resistance) that a car needs to overcome as it is driven on the road. These coefficients are determined through a series of coast-down tests on an outside track. There are a number of aspects of this road load determination procedure that offer vehicle manufacturers potential for exploiting tolerances and flexibilities, ICCT said. These include tire selection and preparation, selection of the test track, ambient test conditions, and pre-conditioning of the vehicle, among others. The ICCT team estimated that about one-quarter of the overall gap observed in 2014 is explained by exploitation of tolerances and flexibilities in the methods required by the EU regulation for determining road load.

• Technology deployment. Certain technologies, such as stop-start systems and hybrid powertrains, have a different effect on CO₂ emissions in the type-approval procedure than they do during real-world driving, because of specific characteristics of the driving cycle used in type-approval testing that differ from typical everyday vehicle operation. ICCT estimated that about one-tenth of the gap in 2014 is explained by an increasing market share of those technologies.

• Other parameters. Running air conditioning systems and entertainment systems increases fuel consumption during real-world driving. Nonetheless, these devices are either switched off or are not fully taken into account during the type-approval emissions test, leading to unrealistically low CO₂ emission values. ICCT put the contribution of these factors at about another one-tenth of the gap in 2014.

For car manufacturers, the divergence puts those that want to report realistic CO₂ emission values at a competitive disadvantage. unrealistic type-approval values also have the potential to damage manufacturers’ credibility and may erode consumer and regulator trust in the entire industry.

—

—“From Laboratory to Road: A 2015 update”

Manufacturers measure vehicle fuel consumption in a controlled laboratory environment, using a test procedure called the New European Driving Cycle (NEDC). This procedure was developed in the 1980s and was not originally intended to be used for fuel consumption testing. A new and more appropriate test procedure, the Worldwide Harmonized Light Vehicles Test Procedure (WLTP), has been developed through the United Nations and is ready for implementation in the EU as early as 2017.

Without introduction of the new test procedure, the WLTP, we expect that the gap will grow to about 50 per cent by 2020. The WLTP would cut the gap approximately in half, and should therefore be introduced in the EU by 2017.

—Dr. Peter Mock, Managing Director of ICCT Europe

The ICCT researchers warn that the WLTP will not resolve all open issues, and could itself inadvertently contain new loopholes that could lead to the performance gap to increase again in the future. Further actions are therefore required, in particular compliance testing of randomly selected production vehicles, they suggested.

Resources

• Uwe Tietge, Nikiforos Zacharof, Peter Mock, Vicente Franco, John German, Anup Bandivadekar (ICCT), Norbert Ligterink (TNO), Udo Lambrecht (IFEU) (2015) From laboratory to road: A 2015 update